Shop OBEX P1 Docs P2 Docs Learn Events
Selecting the right flyback diode when using PWM on 12V 3A coil? — Parallax Forums

Selecting the right flyback diode when using PWM on 12V 3A coil?

eagletalontimeagletalontim Posts: 1,399
edited 2013-10-26 10:13 in General Discussion
I have been all over the place on the web trying to figure out what would be the best flyback diode to use in my circuit. Currently, I have been using the Mosfet's internal diode protection with 500ma coils and it has handled quite well for about a year on the same setup, but I am wanting to beef things up a bit and run a 3A coil at about 45 Hz. If at all possible, I would like to go surface mount to conserve space, but if I need to go through hole, I must make sure it is small enough to fit 4 of them in a 1/2" square.

Comments

  • Mike GreenMike Green Posts: 23,101
    edited 2013-10-24 22:11
    With through hole, you can usually stand them on end. Pretty much any silicon switching diode will do. You can use a 1N914 or 1N4148 for sure. The flyback diode has to be able to handle the current produced by the collapsing magnetic field of the coil which is much much less than the current driving the coil. The reverse voltage it has to withstand is just the voltage driving the coil (12V) plus a fudge factor and the diodes mentioned are good for at least 75V and 300mA.
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited 2013-10-25 06:34
    Hi Mike;
    Mike Green wrote: »
    With through hole, you can usually stand them on end. Pretty much any silicon switching diode will do. You can use a 1N914 or 1N4148 for sure. The flyback diode has to be able to handle the current produced by the collapsing magnetic field of the coil which is much much less than the current driving the coil. The reverse voltage it has to withstand is just the voltage driving the coil (12V) plus a fudge factor and the diodes mentioned are good for at least 75V and 300mA.
    I do not agree with the use of small signal diodes in relay flyback applications such as this. See this wiki:
    Flyback Diodes
    Since eagle has a relay coil rated for 3A the diode must be able to conduct 3A at least for a few mS.
    1N914/1N4148 diodes have a repetitive forward current of 400mA max.
    1N400X 1A rectifier diodes have a peak current rating of 30A. Clearly up to the job.

    Since eagle wants to operate the relay 45 times per second a proper snubber circuit that collapses the coil current faster than the simple single diode snubber is needed.
    I suggest the use of a diode in series with a resistor. This is a bit complicated to calculate.
    Assumptions:
    Supply voltage = 12V
    MOSFET rated at 30V, although prudent design would limit the voltage to 25V
    3A coil implies a coil resistance of 12V/3A=4Ω
    Snubber Diode voltage = 1.1V see spec.
    Snubber Resistor = (25V-12V-1.1V)/3A=3.96Ω
    I can't calculate the field collapse time as the inductance of the coil is not known.
    The field collapse time would be about 1/2 the field charging time as the total resistance of 8Ω is about double the coil resistance of 4Ω.

    If a higher voltage MOSFET is used:
    MOSFET rated at 60V, although prudent design would limit the voltage to 50V
    Snubber Resistor = (50V-12V-1.1V)/3A=12.3Ω
    The field collapse time would be about 1/4 the field charging time as the total resistance of 16Ω is about 4 times the coil resistance of 4Ω.

    Duane J
  • eagletalontimeagletalontim Posts: 1,399
    edited 2013-10-25 06:48
    Would a 1/4 watt resistor be able to handle the current for the short periods the coil will be run at 45 Hz? It should only be ran for about 1.5 seconds every 2 to 3 seconds. I am limited on space, and was hoping a single diode would work. The attached image is a screen shot of my circuit board. Hopefully it looks ok as this is only my second time designing a circuit board! The area highlighted in red is the area I can fit the snubber circuit. I need 1 snubber and 3 standard flyback as those coils are just being switch on or off.
    mainboardv3.jpg
    858 x 818 - 230K
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited 2013-10-25 07:09
    Hi eagle;

    I can't calculate the snubber resistor power dissipation as the coil inductance is not known.
    However, I suspect 1/4W will be OK.

    Where is the relay located on your board?
    What MOSFET are you using and where is it located?

    Duane J
  • eagletalontimeagletalontim Posts: 1,399
    edited 2013-10-25 07:25
    The mosfets are the 2 chips located in the upper area of the red circle. Those are these : http://www.digikey.com/product-detail/en/BTS5210G/BTS5210GINCT-ND/1557902

    I know I will be pushing them a bit with a coil that needs "3A", but the actual ohm reading of the factory coil is 7.8 ohm and a 10 watt 18K resistor is placed in series with the coil for some reason. Still tying to learn about that. I am calculating for a worst case scenario of 16V and a 7.8 ohm load

    The coils are located in the vehicle and will just be controlled by my circuit. I am using the above Profet (not mosfet) to control the common grounded solenoids.

    EDIT * I said that wrong when I said Mosfet... it is actually a Profet....

    Also... These come with a built in load dump protection, but I just want to play it safe. If this additional circuitry it is not needed, I am more than glad to not use it :p
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited 2013-10-25 08:13
    Hi Eagle;
    eagletalontim wrote: »
    I know I will be pushing them a bit with a coil that needs "3A", but the actual ohm reading of the factory coil is 7.8 ohm and a 10 watt 18K resistor is placed in series with the coil for some reason. Still tying to learn about that. I am calculating for a worst case scenario of 16V and a 7.8 ohm load
    Hmmm, can't be 18K. 10W and 18K implies over 400V

    Duane J
  • eagletalontimeagletalontim Posts: 1,399
    edited 2013-10-25 08:18
    On the factory controller, it us a large resistor that says 18(ohm symbol)K and it has 10H marked above that.
  • eagletalontimeagletalontim Posts: 1,399
    edited 2013-10-25 08:30
    Ok, maybe i was wrong. The resistor is actually going in from one Mosfet? or transistor? to another one. I would guess it would be from the Emitter to the base?
  • eagletalontimeagletalontim Posts: 1,399
    edited 2013-10-25 11:34
    I just ran some tests with the Profet I am using and it handled a 2 amp coil at ~45hz for 1 minute with no external flyback diode. There were no ill effects and it did not get warm. I think I may be good without the snubber circuit.....unless it is recommended?
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited 2013-10-25 13:17
    Hi eagle;
    I read the spec sheet.
    While I don't quite understand everything, it appears the device actively clamps the output to zero volts, or so, when turning of inductive loads. It does this by modulating the high side MOSFET in a way to absorb the kickback current.

    A very interesting, if a bit expensive, high side driver that should work well in your application.
    It appears you don't need the snubber diodes.

    Duane J
  • pjvpjv Posts: 1,903
    edited 2013-10-25 15:44
    Mike (and Eagle);

    What Duane said about the current through a flyback diode is absolutely correct. The initial current through the diode is in fact the magnitude of the coil current at the time of switch-off. Remeber, the coil's current cannot change instantaneously, and wants to contnue flowing. The diode is its only path. The current reduces according to the dissipation of the coil's energy into the diode and other associated loop resistance. In theory, at least, with a perfect coil of zero resistance, and a diode with zero forward voltage and zero resistance, the current would flow forever.

    Cheers,

    Peter (pjv)
  • Mark_TMark_T Posts: 1,981
    edited 2013-10-26 01:53
    Mike Green wrote: »
    With through hole, you can usually stand them on end. Pretty much any silicon switching diode will do. You can use a 1N914 or 1N4148 for sure. The flyback diode has to be able to handle the current produced by the collapsing magnetic field of the coil which is much much less than the current driving the coil. The reverse voltage it has to withstand is just the voltage driving the coil (12V) plus a fudge factor and the diodes mentioned are good for at least 75V and 300mA.

    No, the flyback diode takes over the full forward current through the coil which then decays due to losses in the forward voltage
    drop of the diode and resistive losses in the coil itself. So a diode rated at 3A peak is needed and a 1N4148 will just fuse.

    Remember inductors resist change in current so with 3A flowing and you switch off that 3A still tries to flow, and it finds the
    path of least resistance (literally), through the diode (rather than arcing across a switch contacts or avalanche breakdown
    of a switching transistor). V dt = - L dI
  • Alexander (Sandy) HapgoodAlexander (Sandy) Hapgood Posts: 360
    edited 2013-10-26 10:13
    Littelfuse make special TVS diodes. They have some technical information on their site that is worth a read.

    Turns out that a simple diode isn't the best for suppressing the voltage spikes from a relay because it can slow the relay down and cause arching of the relay contacts. Littelfuse have products that are essentially back to back Zener diodes that speed up relay operation and are not polarity sensitive. They also have high current capacity.
Sign In or Register to comment.